Temporally robust models for predicting seed yield of moist‐soil plants

Abstract

ABSTRACT Rapid assessment of food production and subsequent availability is fundamental to evaluating wetland management practices and general habitat quality for waterfowl. Traditional methods of estimating food biomass (e.g., plot and core sampling) require considerable time, expertise, and cost. Rapid assessment models using plant measurements or scanned seed‐head area have recently been adapted to predict seed production in moist‐soil wetlands. We evaluated existing models of seed production and estimated benthic seed density with data collected during autumn 2011 in western Tennessee, USA, to improve prediction capability of seed availability for waterfowl. Generally, all models explained significant variation (r2 = 0.85–0.98) and accurately predicted seed production in moist‐soil plants (r2 = 0.84–0.97). Belowground proportions of seed biomass and duck energy days differed across species relative to previously reported biomass estimates in moist‐soil wetlands (x¯ = 0.4–9.1%); thus, production estimates from models should be adjusted on a species‐specific basis and the effect of belowground seeds on overall energetic carrying capacity estimates will vary with species composition of wetlands. We recommend use of updated most‐soil rapid‐assessment models incorporating seed bank estimates to predict waterfowl food availability and evaluate management practices. © 2017 The Wildlife Society.

Keywords

• carrying capacity
• dabbling duck
• Tennessee National Wildlife Refuge
• waterfowl
• wetlands